| # Goodreads Datasets | |
| ## Dataset Description | |
| The Goodreads datasets consist of four datasets, specifically labeled as Goodreads-History, Goodreads-Crime, Goodreads-Children, and Goodreads-Cosmics. The Goodreads datasets are a user-book review network. It includes information about books, users and reviews. Nodes represent books and users. Text on a book node is the description of the book. Text on a user node is the `user`. The book text includes the following information: `The book [title] is a [format] edition published by [publisher] in [publication_month] [publication_year] about [description], consisting of [num_pages].` Edges represent relationships between books and users. Text on an edge means a user's review of a book. | |
| ## Graph Machine Learning Tasks | |
| ### Link Prediction | |
| Link prediction in the Goodreads dataset involves predicting potential connections between users and books. The goal is to predict whether a user will review a book. | |
| ### Node Classification | |
| Node classification tasks in the Goodreads dataset include predicting the book's category. | |
| ## Dataset Source | |
| https://mengtingwan.github.io/data/goodreads.html |